Comparing Calculated LDL with Direct Method: A Retrospective, Real-World Evidence Study on Diagnostic Lab Reports from A Single Center in India

Abstract

Objective: To compare and validate the calculated LDL values from Friedewald and Anandaraja formulas with directly measured values in the Indian population. Material and Methods: The study was conducted on randomly selected 102 individuals of 16 to 88 years of age during December 2019. The direct LDL values were measured using selective solubilization assay, and Friedewald and Anandaraja formulas were used to calculate LDL for comparison. The correlations between direct and calculated methods were assessed using the linear regression method. Receiver operating characteristic analysis with nonparametric distribution was used to compare the sensitivity and specificity of the three methods. Results: The average LDL of direct method, 107.3 mg/dL, Friedewald formula, 89.7 mg/dL, and Anandaraja was 99.0 mg/dL. The relation between direct and calculated values assessed by linear regression showed 97% and 87% of correlation with Friedewald and Anandaraja, respectively. The ROC analysis inferred that direct (AUC 0.74; 95% CI 0.64-0.83) and Friedewald (AUC 0.71; 95% CI 0.61-0.81) methods had shown about 70% efficiency in predicting true positive and true negative dyslipidemia cases. In our dataset, the Anandaraja formula could not well differentiate positives from negative cases of dyslipidemia with merely 60% AUC. Conclusion: The underpredicted values from the Friedewald formula were associated with deranged cholesterol and HDL values, not triglycerides. Anandaraja formula overpredicted by 10 to 30 mg/dL when triglycerides were <150 mg/dL and underpredicted by 10-43 mg/dL while non-HDL was >140 mg/dL.